A predictive model for determination of sand liquefaction potential based on energy method

The prediction of liquefaction potential of sand under earthquake action is a complicated problem in civil en-gineering. In this study, gene expression programming (GEP) was used to establish the relationship between the capacity energy required to trigger sand liquefaction and several major parameters. By collecting 85 experi-mental data and six main parameters, a GEP model was established and the empirical prediction equation was obtained. Three statistical indices were used to evaluate the performance of the GEP model and 11 other existing capacity energy models. The results show that the coefficient of determination (R2), mean absolute error (MAE) and root mean squared error (RMSE) values of all datasets predicted by the GEP model are 0.924, 0.061 and 0.087, respectively. While the R2, MAE and RMSE values of all datasets of the 11 existing capacity energy models are 0.510-0.777, 0.149-0.468 and 0.196-0.653, respectively. It can be concluded that the proposed GEP model has the potential to accurately predict the capacity energy required to trigger sand liquefaction.

Automated summary

In this study, gene expression programming (GEP) was used to establish the relationship between the capacity energy required to trigger sand liquefaction and several major parameters. The GEP model showed higher accuracy and better performance compared to existing models, as confirmed by experimental data.